ABSTRACT
Psychometric tests are composed of individual items, and the common characteristics of tests can more often than not be applied to individual items as well. Thus the simple question‘Are you interested in this job?’ can be scored (1 for yes and 0 for no, or vice versa), can be unreliable (that is, some people may give different answers each time), can be invalid (the answers may be wrong), or it can be biased (such as when some types of people may be more likely to lie than others). It can also be administered by interview, it can be judged by observation, or it can be tested by a paper and pencil multiple-choice item (tick one box for‘yes’, the other for‘no’). Although there are some cases where selection might be made purely on the basis of one question (for example‘What GCSEs do you have?’), this is unusual. More often there are many aspects we might wish to cover, especially as it is usually the case that the opportunity to ask one question provides an opportunity to ask many more. However, all tests are composed of individual items as their elements and so the success of a test depends on the success of its items.
